To Make or not to Make: Using a model-based parameter to detect vegetation changes from polarimetric and interferometric radar data
This project focuses on implementing a software component that will help the NASA JPL (Jet Propulsion Laboratory) UAVSAR (Uninhabited Ariel Vehicle Synthetic Aperture Radar) group ultimately estimate Earth’s biomass and vegetation properties. Estimating the worldwide biomass will contribute to our understanding of the changing climate. By utilizing Cloude’s model-based incoherent decomposition we are able to extract the maximum surface-to-volume scattering ratio from the polarimetric and interferometric SAR (Pol-InSAR) data. The surface-to-volume ratio contains information about the structure of the imaged vegetated area, and can be used to detect vegetation changes between data acquisitions. The software component will create an image mask using a threshold based on the surface-to-volume ratio. The mask will be needed at times when vegetation changes between data acquisitions are too large to allow for accurate estimation of forest parameters from Pol-InSAR data. In order to implement the mask, we implemented the mathematical model in C programming language and incorporated it into the parameter estimation program. In the presentation we will describe the theoretical model and show results of the estimation surface-to-volume ratio on using UAVSAR collected over the Harvard Forest (MA).